El gobierno de Allende y la comunicación internacional

Quality control tests of the finished filter can be divided into two categories: direct and indirect tests. Direct tests evaluate microbiological removal efficiency. Due to the associated costs, microbiological testing may be carried out on a sample of filters; however,

manufacturing consistency must be verified in order to establish that the samples are representative. Indirect tests are carried out to evaluate manufacturing consistency. Indirect tests should be carried out on 100% of the filters considered for consumer use.

Indirect tests include visual inspection, auditory testing, pressure testing, and flow rate testing. In addition, the filter rim should be checked to ensure that it fits the receptacle properly. These evaluations should be conducted before silver application so that neither silver nor application time are wasted on filters that will potentially be rejected by quality control and to minimize the amount of silver discarded into the environment.

Factories should strive to achieve a consistent rejection rate of 10%, although 15-20% can also be considered acceptable. Although manufacturing cost is an important criterion for defining acceptable failure rates, a rejection rate of more than 20% or a change in the average rejection rate indicate inconsistency in production practices, materials, or methods and should not be considered acceptable. Conversely, a rejection rate of less than 5% likely indicates that quality control evaluations are not strict enough.

8.3.1 Visual Inspections

Visual inspections should take place before each major step of the production process so defective filters can be removed from the production line. Formal visual

inspections should be carried out and documented before: 1) surface finishing; 2) loading the kiln; 3) flow rate testing; 4) silver application; and, 5) packaging. The documentation of each formal visual inspection, including the production stage and description of the defect, will help when troubleshooting a quality control concern and will also encourage employees to give their full attention to the visual inspection process.

In the greenware state, filters should be examined for cracks, warping, inconsistent filter walls, large pieces of burn-out material, and inconsistent surface. In fired filters, filters should be examined for discoloration (including blackened areas indicating insufficient oxidation during firing), warping, cracks, holes or spaces from large pieces of burn-out material, charring, and crumbling. The angle between the rim and wall of the filter and between the base and wall should also be checked for proper alignment.

The filter rim of fired filters should be checked for size and warping by placing a receptacle lid on each filter element. Turn the lid slowly and check that the filter rim meets the lid evenly. If the lid does not fully cover the filter rim, grind the filter rim to fit, taking care to not damage the body of the filter or grind any more material than necessary. The filter rim must still be wide enough to cover the rim of the fitting ring or receptacle, so that untreated water or other debris cannot enter the filter during use; otherwise the filter must be rejected.

At the Indo-2 factory, to emphasise and reinforce the importance of filter inspections,

two people perform visual inspections of finished filters

using a magnifying glass; defects are marked with chalk. Defective filters are compared

and analyzed to establish the cause of the defect.

8.3.2 Auditory Testing

Auditory testing can identify incomplete firing or cracks in the filter walls. After being fired, each filter should be tapped to check for a ringing sound. A suggested test method is as follows:

1) balance the filter by its base on an open hand or outspread fingers;

2) tap or knock near the rim of the filter, as if ringing a bell;

3) if dull thud is heard, the filter should be rejected;

4) tap the filter in at least three locations around the rim of the filter.

A properly fired filter will make a ringing sound. A filter that makes a blunted, shorter sound, almost a thud, is cracked or improperly fired and should be rejected.

In addition, the open end of the filter can be held up to one’s ear and the filter walls gently squeezed. A slight crunching sound will be heard if there are cracks present, and that filter must be rejected.

8.3.3 Pressure Testing

Before saturating filters for flow rate testing, a pressure test should be carried out on all fired filters to check for cracks or large pores that allow water to pass through the filter walls too quickly. This test should be performed as follows:

1) hold the filter by the rim; 2) submerge the filter base in

water until the water level is near the rim. Do not allow water to flow into the filter (Figure 8-1);

3) keep the filter submerged for 10 seconds;

4) if evidence of water entering the filter (Figure 8-1) is

present after 10 seconds, especially if it seeps in unevenly, the filter has cracks or large pores and must be destroyed.

Another way to carry out the pressure test is to invert the filter and fully submerge it rim down in water, trapping air inside the filter. The presence of large pores or cracks may be identified by a stream of bubbles coming from the wall or base of the filter. If a stream of bubbles appears, the filter must be destroyed.

In Myanmar, third party inspections are performed

on a random selection of 5% of all filters sold to NGOs. If any filters fail inspection, the entire shipment is held up until

the issue is resolved.

Figure 8-1: Pressure Test and Evidence of Leaks (Pillers and Diaz 2009)

8.3.4 Flow Rate

The flow rate is the amount of water that passes through a full, saturated filter in the first hour. It is used as an indicator of production consistency, potential contact time with silver, and can also identify cracks or large pores in the filtering element.

The minimum flow rate is established based on consumer needs. A minimum flow rate of one liter in the first hour should provide enough drinking water for a family of five if

the filter is filled 4-5 times a day. Filters that do not meet a minimum flow rate of one liter per hour should be rejected and destroyed since their failure to treat a sufficient quantity of water could cause the consumer to stop using the filter, thereby placing their health at risk. Filters that do not meet the minimum flow rate may not have reached the appropriate peak

temperature during firing, so these filters can be refired. Often this will bring a filter into compliance, but if after a second firing, a filter still does not fall within the flow rate range, it should be destroyed.

Since the amount of time it takes water to pass through the filter will influence some mechanisms of filtration, the thickness of the filter walls and the surface area of the filter should be taken into consideration when determining the maximum flow rate for each filter design. Assuming a minimum 2 cm base thickness and 1.5 cm wall thickness in flat-

bottomed filters or 1.5 cm wall thickness in rounded-bottomed filters, the maximum flow rate should be calculated at 0.35 liters per hour per liter capacity of the filter element. Therefore, the maximum flow rate for 7.2-liter capacity filters should be 2.5 liters in the first hour, and a 10-liter capacity filter should have a flow rate of no more than 3.5 liters. Maximum flow rates should be confirmed with microbiological testing and filters that exceed the acceptable flow rate should be destroyed.

While the hydraulic conductivity will influence filtration mechanisms and treatment

effectiveness, the flow rate does not guarantee microbiological efficacy. When a factory is first started, prototype filters that meet the flow rate requirements are manufactured and tested for microbiological effectiveness. Since there are many production variables that can influence the flow rate (Section 5.2), filters must be produced with similar materials and methods and fall within the flow rate range in order to be considered representative of the prototype filters. When materials or production vary, microbiological effectiveness must be confirmed, as during prototype evaluations. If there is a high variation in flow rate within a given batch, indicating inconsistency in production, that batch must be held from distribution or sale until microbiological efficacy is confirmed and the materials and production of that batch are evaluated.

8.3.4.1 Establishing Saturation Time

For reliable results, filters should be saturated before flow rate testing (Lantagne 2001a). Although the amount of time required to ensure saturation might vary per factory, Nederstigt and Lam (2005) found that after having soaked for 24 hours, filtration rates became

constant. Filters soaked for fewer than 24 hours had lower flow rates (van Halem 2006: A- 10). Water quality requirements for soaking filters are discussed in Section 4.1.

In order to establish the amount of time it takes to saturate the filters to achieve a consistent flow rate, a number of filters that have been taken from different locations in the kiln should

The Myanmar factories report that filters (10 L capacity) with flow rates of 3.5 and 4.5 liters in the first hour have the

same likelihood of achieving a 2-log reduction of indicator

be fully submerged. The time the filters are submerged should be recorded. At regular intervals, filters should be removed from the water and their flow rate tested, and the filters returned to the soak tank. When the filtration rate of a specific filter is consistent for two or three consecutive readings, that filter can be considered saturated for flow rate testing purposes. The maximum time it takes the sampled filters to become saturated can be used as the minimum time necessary for saturating filters for future flow rate testing. Saturation time can be expected to be at least 12 hours, and as much, or more than, 24 hours. Based on variations in materials and manufacturing, a different saturation time may be required for each factory; however, filters should be soaked for a consistent amount of time at each factory. This saturation test should be carried out periodically to confirm that filters continue to be fully saturated before flow rate testing.

8.3.4.2 Flow Rate Testing

The flow rate can be tested by measuring the amount of water filtered after 1-hour. With flat bottomed filters, rather than measuring the amount of water filtered (effluent), the drop in water level can be measured using a calibrated T-device (Figure 8-2). Instructions for making a T-piece are in Annex I. Since the drop in water volume will result in a reduced hydraulic head, the filtration rate will slow as the filter empties; therefore, readings taken after half an hour and then doubled may not provide accurate readings. Water quality requirements for flow rate testing are discussed in Section 4.1.

Saturated filters should be placed in an empty bucket or receptacle supported by the rim only. Flat-bottomed filters should not be resting on their base, as this will reduce the flow rate. Likewise, if the water level in the receptacle

reaches the bottom of the filter, the flow rate results will not be accurate. Therefore, if the maximum allowed flow rate is 2 liters in the first hour, there should be room for more than 2 liters of water in the receptacle so that, within the acceptable flow rate range, the water level does not reach the base of the filter element.

Filters should be completely filled with water and allowed to filter for 1 hour. If measuring the effluent, carefully remove the filter so that the water remaining in the filter does not spill into the receptacle. Measure the quantity of water that passed through the filter into

the receptacle. Alternatively, if it is possible to see the water level through the bucket, lines can be drawn on the bucket at half-liter intervals; however, the receptacle must be on a level surface for this to be accurate. When using a calibrated T-device, the drop in water level can be measured after 1-hour.

Flow rates should be recorded on the filter production log. Average flow rates should be monitored throughout production. Filter elements that fall within the acceptable flow rate range can be emptied and left to dry before silver application and packaging. However, if there is a large variation in flow rates for filters in a given batch, the production records for that batch should be reviewed and compared with the records and results from other batches to see if the cause(s) can be identified, additional microbiological testing should be carried out, and that batch may need to be withheld from distribution until causes are identified and efficacy confirmed.

Figure 8-2: Calibrated T-device (source unknown)

8.3.5 Microbiological Testing

Each factory should establish a testing protocol for both independent laboratory and in-house testing which includes: frequency of testing; percentage of filters from each batch tested; the method for selecting filters; indicator used; and, testing methods.

At a minimum, microbiological efficacy must be confirmed by independent laboratory testing on at least three prototype filters

from three different filter batches (a total of nine filters) before production starts, when the mixture ratio or other aspect of production varies, or when a quality control issue arises. Throughout production, for continued verification of production methods, a minimum of 0.1% of filters should be tested at an independent laboratory. When taking samples to a laboratory for testing, the laboratory should be contacted in advance for requirements and guidelines for sample collection. In addition, at least one filter per firing batch and a minimum of 1% of filters destined for sale (that have passed all other quality control inspections) should be tested in-house, at the factory.

If possible, more filters should be tested, the same filters should be tested more than once to confirm results, and an increased number of filters should be tested if there is a change in production or a change in the quality or consistency of filters produced.

Filters selected for testing should be representative of their batch; therefore, they should be selected from different drying locations, different locations in the kiln, from those

manufactured during different shifts during the work day, and so on. Microbiological testing should be carried out on filters that have already passed other quality control inspections and tests (Section 8.3), but before applying silver. This avoids wasting money and time testing noncompliant filters, prevents the initial high silver concentration in the filter effluent from influencing test results, and finally, because the duration of the efficacy of silver is unknown, testing without silver provides a more reliable indication of long-term performance. Testing can be carried out on filters with silver if silver has been fired into the filters or if the objective of the test is to measure silver in filter effluent.

If a filter that has met all other quality control tests does not pass microbiological tests, the entire batch must be held from distribution until additional filters from that batch have been tested. The production records of that batch of filters should be evaluated and the problem(s) diagnosed and resolved. Following this, an increased number of filters from subsequent batches must be tested in order to confirm consistency in production. If filters consistently demonstrate microbiological effectiveness, the number of filters tested at an independent laboratory can be reduced, as long as all

other aspects of production remain consistent and in-house testing continues. However, routine verification by an independent laboratory should never be eliminated and the frequency of testing should be increased again if any inconsistencies are noticed.

The fact sheet Microbiological Indicator Testing in Developing Countries: A Fact Sheet for

the Field Practitioner explains the different indicator bacteria and their usefulness; currently

available and emerging field test methods including the benefits, drawbacks, and cost

At the FilterPure factory in the DR, no changes in production procedures are allowed until efficacy

is confirmed by microbiological testing.

In Myanmar, a traveling technician with a field test kit visits each of the eight factories

once per month and tests 10- 15 randomly selected filters selected. The technician spends the night and shows

the filter manufacturers the results in the morning.

considerations; and, detailed instructions for carrying out the tests. A copy of this document can be obtained by contacting Daniele Lantagne at: [email protected]. Much of the following sections have been summarized from this fact sheet.

8.3.5.1 Indicator Organisms

Due to the difficulty of monitoring water for specific contaminants, water is tested for the presence of indicator organisms associated with fecal contamination. Commonly used indicator organisms include: 1) total coliforms (TC); 2) thermotolerant or fecal coliforms (TTC); and, 3) E. coli. Criteria for indicator bacteria as outlined by the WHO (2006: 142) are

that they should be:

1) universally present in high numbers in human or other warm-blooded animal feces; 2) readily detectable by simple methods; and,

3) should not grow in natural water.

Since some coliforms can grow and survive in water and are often present in the absence of fecal contamination, the TC count is not always useful in evaluating health risk; however, it can be useful as an indicator of treatment effectiveness (WHO 2006: 283). TTC are those of the TC group that are able to ferment lactose at 44-45°C. Some TTC are present in the natural environment in tropical countries. E. coli, often the predominant TTC organism, is rarely found in the absence of fecal contamination and is therefore a more reliable indicator of water safety, but both E. coli and TTC are considered acceptable indicators to measure water safety (WHO 2006: 284). Another indicator, hydrogen sulfide-producing bacteria, is not recommended for testing filter efficacy as currently there is no quantitative testing method available.

The choice of indicator bacteria will depend upon the quality of the challenge water and whether the safety of the water or the treatment efficacy is being tested. Since TC and TTC are more likely to be present in greater numbers, they are often used in place of E. coli to

evaluate treatment efficacy (percent reduction or LRV). Ideally, filters should be challenged with water containing high

concentrations of E. coli so that both percent reduction and the safety of the water can be evaluated.

8.3.5.2 Measuring Microbiological Efficacy

Microbiological efficacy of filters can be measured by percent reduction of indicator bacteria expressed as LRV (Table 1) or by evaluating the risk level of treated water (Table 2). Challenge water must contain the indicator bacteria being tested for in order to test for treatment effectiveness.

To establish LRV, the indicator bacteria must be present in greater numbers and quantified in both the challenge and treated water. For E. coli, TTC, and TC, filters must, at a minimum, provide a 2-log reduction. This means that the effluent concentration of the indicator must be